Automated Hypofractionated IMRT treatment planning for early-stage breast Cancer

Page created by Kirk Knight
 
CONTINUE READING
Automated Hypofractionated IMRT treatment planning for early-stage breast Cancer
Lin et al. Radiation Oncology    (2020) 15:67
https://doi.org/10.1186/s13014-020-1468-9

 RESEARCH                                                                                                                                       Open Access

Automated Hypofractionated IMRT
treatment planning for early-stage breast
Cancer
Ting-Chun Lin1†, Chih-Yuan Lin1†, Kai-Chiun Li1, Jin-Huei Ji1, Ji-An Liang1,2, An-Cheng Shiau1,3,4,
Liang-Chih Liu2,5 and Ti-Hao Wang1*

  Abstract
  Background: Hypofractionated whole-breast irradiation is a standard adjuvant therapy for early-stage breast cancer.
  This study evaluates the plan quality and efficacy of an in-house-developed automated radiotherapy treatment
  planning algorithm for hypofractionated whole-breast radiotherapy.
  Methods: A cohort of 99 node-negative left-sided breast cancer patients completed hypofractionated whole-breast
  irradiation with six-field IMRT for 42.56 Gy in 16 daily fractions from year 2016 to 2018 at a tertiary center were re-
  planned with an in-house-developed algorithm. The automated plan-generating C#-based program is developed in
  a Varian ESAPI research mode. The dose-volume histogram (DVH) and other dosimetric parameters of the
  automated and manual plans were directly compared.
  Results: The average time for generating an autoplan was 5 to 6 min, while the manual planning time ranged from
  1 to 1.5 h. There was only a small difference in both the gantry angles and the collimator angles between the
  autoplans and the manual plans (ranging from 2.2 to 5.3 degrees). Autoplans and manual plans performed similarly
  well in hotspot volume and PTV coverage, with the autoplans performing slightly better in the ipsilateral-lung-
  sparing dose parameters but were inferior in contralateral-breast-sparing. The autoplan dosimetric quality did not
  vary with different breast sizes, but for manual plans, there was worse ipsilateral-lung-sparing (V4Gy) in larger or
  medium-sized breasts than in smaller breasts. Autoplans were generally superior than manual plans in CI (1.24 ±
  0.06 vs. 1.30 ± 0.09, p < 0.01) and MU (1010 ± 46 vs. 1205 ± 187, p < 0.01).
  Conclusions: Our study presents a well-designed standardized fully automated planning algorithm for optimized
  whole-breast radiotherapy treatment plan generation. A large cohort of 99 patients were re-planned and
  retrospectively analyzed. The automated plans demonstrated similar or even better dosimetric quality and efficacy
  in comparison with the manual plans. Our result suggested that the autoplanning algorithm has great clinical
  applicability potential.
  Keywords: Automation, Treatment planning, Autoplanning, Hypofractionation, IMRT, Early-stage, Whole-breast
  irradiation, Left-sided breast cancer

* Correspondence: thothwang@gmail.com
†
 Ting-Chun Lin and Chih-Yuan Lin contributed equally to this work.
1
 Department of Radiation Oncology, China Medical University Hospital, China
Medical University, Taichung, Taiwan
Full list of author information is available at the end of the article

                                       © The Author(s). 2020 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
                                       International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
                                       reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
                                       the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
                                       (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Automated Hypofractionated IMRT treatment planning for early-stage breast Cancer
Lin et al. Radiation Oncology   (2020) 15:67                                                                    Page 2 of 9

Background                                                    patient manual treatment plans who received hypofrac-
Breast cancer is the second most prevalent cancer and         tionated whole-breast irradiation as an adjuvant therapy
the second leading cause of cancer death among women          for partial mastectomy at our hospital to prove this con-
worldwide, with a slight but stable increase in its inci-     cept. These treatment plans were re-planned with an in-
dence in the most recent decade [1]. At our center,           house-developed algorithm. The dosimetric quality was
breast cancer patients account for approximately one-         compared between the automated plans and manual
third of all patients requiring curative radiotherapy. For-   plans.
tunately, most patients are diagnosed at early stage with-
out nodal involvement, and for this subgroup, the             Methods
standard treatment consists of surgery and postoperative      Patient selection
radiotherapy, and systemic adjuvant therapy if necessary.     A cohort of 99 patients diagnosed with left-sided stage I
Postoperative whole-breast irradiation serves as an adju-     or stage II node-negative breast cancer who received
vant therapy following breast-conserving surgery that         postoperative whole-breast irradiation without nodal ir-
provides equivalent long-term survival comparable to          radiation from year 2016 to 2018 at a tertiary center
radical mastectomy, and is now recognized as the stand-       were included in this study. All patients were aged
ard treatment for early-stage breast cancer [2]. For          20 years or older and received hypofractionated regimen
node-negative patients, the hypofractionated schedule is      of 42.56 Gy in 16 daily fractions.
commonly recommended as randomized trials have con-
firmed their safety and efficacy [3]. At our department,      CT simulation and contouring
hypofractionated radiotherapy (42.56 Gy in 16 fractions)      A customized immobilization device was used for all pa-
has been implemented for nearly all early-stage node-         tients with the ipsilateral arm raised to maximize the
negative breast cancer patients in the recent 2 years.        precision and repeatability of the daily positioning for ir-
   Following the Quantitative Analysis of Normal Tissue       radiation. Following patient immobilization, planning
Effects in the Clinic (QUANTEC) and Radiation Ther-           images were acquired with a computerized tomography
apy Oncology Group (RTOG) consensus guidelines, the           (CT) scanner (SOMATOM Definition AS, Siemens,
main organs at risk (OARs) for whole-breast radiother-        Germany) at a 3-mm slice thickness. Our department
apy include the heart, lung and contralateral breast [4,      used the voluntary deep-inspiration breath-hold (DIBH)
5]. Whole-breast irradiation is generally performed with      technique to reduce the cardiac radiation dose in breast
two opposed tangential beams to avoid the above-              cancer management. The patients performed a super-
mentioned normal organs while simplifying the treat-          vised breath hold during CT simulation and treatment.
ment plan. Tangential beam intensity-modulated radi-             For each patient in our study cohort, we re-planned
ation therapy (IMRT) may be employed to enhance the           with our auto-planning algorithm, based on previous ap-
dose homogeneity [6–9]. Compared with other cancer            proved contours. The treatment target and OARs were
sites, treatment planning for whole-breast irradiation is     manually contoured based on the RTOG 1005 protocol
relatively simple yet contributes to a large proportion of    guidelines [16]. The clinical target volume (CTV) in-
the workload for the medical personnel in radiation on-       cluded the whole left breast and was extended isotropic-
cology departments, and thus is an ideal candidate for        ally with a margin of 5 mm to form the planning target
automation. In the recent decade several published stud-      volume (PTV). The PTV was cropped to a distance of 4
ies reported on automated treatment planning algo-            mm from the patient’s skin surface. The contralateral
rithms implementation for the head and neck, tangent          breast, ipsilateral and contralateral lung, and heart were
breast, prostate and palliative spine [10–15]. These auto-    manually contoured as OARs.
mated planning algorithms reduced the time and effort
required to create personalized treatment plans, while        Manual treatment planning technique
also allowing the planning process to be highly standard-     All manual plans for hypofractionated whole-breast ir-
ized. The published studies generally showed similar tar-     radiation used six-field IMRT design with six MV pho-
get coverage and equivalent clinical acceptability of         ton beams at our department, created by experienced
automated plans when compared with manual plans. For          medical physicists. Two tangential beams (major fields)
head and neck cancer planning, better OAR sparing fa-         were manually assigned by the treatment planner to fit
voring the automated plans were even reported [12].           the PTV borders. Four oblique beams (minor fields) with
   We argue that a well-designed automated planning al-       reduced field size were subsequently created. All of the
gorithm can improve current radiation oncology clinical       fields were extended at the breast apex to account for
workflow by standardizing plan quality, accelerating the      the breathing motion. The minor beam angles were ad-
treatment planning process, and improving employee            justed to avoid the critical organs. Dose-volume con-
productivity. We collected left-sided breast cancer           straints were set based on the RTOG 1005 protocol.
Automated Hypofractionated IMRT treatment planning for early-stage breast Cancer
Lin et al. Radiation Oncology     (2020) 15:67                                                                                        Page 3 of 9

Auxiliary structures Ring_1 and Ring_2 were manually                       gantry angle, iterate between 270 and 330 degrees
created by the treatment planner. The areas of relative                    (300 ± 30 degrees range) for major field F1, and between
overdose (hotspots) were eliminated by assigning the hot                   90 and 150 degrees (120 ± 30 degrees range) for major
spots region, isodose area of V105%, as a constraint struc-                field F2. For the collimator angle, iterate between 330
ture to the objective template. The elimination of hot-                    and 30 degrees (0 ± 30 degrees range). For each iteration,
spots usually needs to be repeated several times. A                        the jaw fits to the PTV and then the field area (defined
beam’s-eye-view example of the manual plan is shown in                     by the X and Y jaws) is calculated. We choose the gantry
Fig. 1.                                                                    and collimator angles pair which has the smallest field
                                                                           area. (4) Add four minor fields F3, F4, F5 and F6. The
Automatic treatment plan (autoplan) generation                             three minor fields F3, F4 and F5 are set based on F1.
The automated plan-generating computer algorithm is                        The gantry angles are set as follows: F3 = F1 + 15 de-
an in-house-developed C# program created as a Script of                    grees, F4 = F1 + 30 degrees, F5 = F1 + 45 degrees, and the
the Varian Eclipse Scripting Application Programming                       collimator angles of these three subfields are set as F1.
Interface (ESAPI). The generated autoplan schematic                        The fourth minor field F6 is determined based on F2,
diagram is shown in Fig. 2. The detailed process is de-                    with F6 gantry angle = F2–15 degrees, and F6 collimator
scribed as follows: (1) Automatically create auxiliary                     angle is set as F2. For the minor fields, the jaws fit to the
structures Ring_1 and Ring_2 by extending the PTV                          PTV, and then the inner side of the jaw opens 1.5 cm
posteriorly (Fig. 1), and then create a new plan under a                   away from the isocenter (Fig. 1). (5) Optimize plan ac-
new course for a selected patient. (2) Set the centroid of                 cording to the objective template (Additional file 1:
the PTV as the treatment isocenter. (3) Identify the opti-                 Table S1) and calculate dose. (6) Evaluate and minimize
mal opposed tangential gantry angles and collimator an-                    V105% hotspot area. If V105% < 0.5 cc, the plan is com-
gles by iterating the angles one degree at a time. For the                 pleted (Step 8). If V105% > 0.5 cc, the constraint structure

 Fig. 1 A representative CT axial view demonstrating beam arrangement and planning auxiliary structure (left panel) and beam’s-eye-view (right
 panel). Upper row: manual plan. Lower row: autoplan
Lin et al. Radiation Oncology     (2020) 15:67                                                                              Page 4 of 9

 Fig. 2 The schematic diagram of an autoplan. Detailed description is in the Methods and Materials section

with region of 105% isodose area will be automatically                     between OAR parameters of patients with different
created, and the program moves to the next step (Step                      breast sizes stratified into three groups (small, medium
7). (7) The V105% constraint structure is added to the ob-                 and large) according to the CTV with a cutoff of 300 ml
jective template (Additional file 1: Table S1). Plan                       and 600 ml. A p-value less than 0.05 was considered sta-
optimization and dose calculation are repeated once                        tistically significant. Data are presented as mean value ±
again. (8) The autoplan is completed.                                      standard deviation (SD) if not otherwise specified. All
  All automated plans were generated using the Eclipse                     statistical analyses and graphs were performed using the
treatment planning system (TPS) under a research li-                       statistical software R version 3.5.2.
cense (Version 15.6, Varian Medical Systems Inc., Palo
Alto, California, USA). Analytical Anisotropic Algorithm                   Results
was used for dose calculation for all plans. All auto-                     The mean difference between the gantry and collimator
mated plans were simulated on the TPS but not deliv-                       angles for F1 and F2 fields between autoplans and man-
ered clinically.                                                           ual plans were 2.69 ± 3.00, 5.24 ± 5.63, 2.26 ± 2.67 and
                                                                           2.42 ± 2.81 (Mean ± SD) degrees, respectively. For hot-
Plan evaluation and analysis                                               spot V105% evaluation of all 99 autoplans, 66 (67.7%)
The DVHs and dosimetric parameters of all manual                           plans had smaller V105% after the automatic optimization
plans and autoplans were collected and calculated for                      step (Step 6), 24 (24.2%) did not have V105% > 0.5 cc, and
data analysis. The mean DVH band was calculated and                        eight (8.1%) plans still had V105% > 0.5 cc after
reflected the volume-dose distribution with 95% confi-                     optimization.
dence interval. For plan evaluation, the following param-                    Table 1 listed the dosimetric parameters for the hot-
eters were recorded. The V110% of the prescribed dose                      spot area (V110%), the PTV coverage, ipsilateral lung,
inside the body was analyzed for hotspot evaluation. For                   heart and contralateral breast. There was no significant
PTV coverage evaluation, V95% was collected. The dose                      difference in hotspot volume between the two ap-
homogeneity index (HI) of PTV was measured by D5%                          proaches. However, compared with the manual plan, the
divided by D95% (D5% / D95%) [17]. The conformity index                    autoplan performed slightly better in ipsilateral-lung-
(CI) of PTV was defined as BV95% / PTV (BV95% = the                        sparing (in terms of V16Gy, 14.0 ± 3.6 vs. 13.3 ± 3.0, p <
volume of the body receiving 95% of the prescribed                         0.01, and mean dose, 6.6 ± 1.5 vs.6.3 ± 1.2, p < 0.01), but
dose) [18]. For dose analysis of the OARs, the V16Gy and                   was inferior in contralateral-breast-sparing (in terms of
Dmean were reported for the ipsilateral lung, V20Gy and                    maximal dose, 3.5 ± 7.3 vs.5.0 ± 8.7, p < 0.01). Addition-
Dmean for the heart, and V5Gy and Dmax for the contra-                     ally, the autoplan was generally superior to the manual
lateral breast per RTOG 1005 protocol. The monitor                         plan in CI (1.24 ± 0.06 vs. 1.30 ± 0.09, p < 0.01) and MU
unit (MU) was calculated by summing the MUs of all                         (1010 ± 46 vs. 1205 ± 187, p < 0.01).
fields in a treatment plan.                                                  The average time to generate an autoplan was only five
                                                                           to 6 min irrespective of each patient’s anatomical vari-
Statistical analysis                                                       ance, while the manual planning time ranged from 1 h
The dosimetric parameters and total MUs were evalu-                        to one and a half hours. Figure 3 showed the autoplan
ated by paired two-tailed t-test for the manual and auto-                  and manual plan dose distributions for patients with dif-
matic treatment plans. One-way analysis of variance                        ferent breast sizes (large, medium and small). Table 2
(ANOVA) was applied to evaluate the differences                            shows the ipsilateral lung, contralateral breast and heart
Lin et al. Radiation Oncology   (2020) 15:67                                                                           Page 5 of 9

Table 1 The mean, minimal and maximal value of the parameters for evaluation of PTV coverage and OARs constraints of all plans.
Data analysis for comparison between manual plans and autoplans was done with paired two-tailed t-test. A p-value less than 0.05
was considered statistically significant
                                 Manual                                      Auto
Parameter                        Mean value ± SD [Min, Max]                  Mean value ± SD [Min, Max]                    p-value
Body
     V110%[%]                    0.07 ± 0.29 [0.0, 2.15]                     0.03 ± 0.23 [0.0, 2.15]                       0.29
PTV
     V95%[%]                     98.5 ± 0.5 [96.7, 99.5]                     98.5 ± 0.6 [96.4, 99.9]                       0.40
Ipsilateral lung
     V16Gy[%]                    14.0 ± 3.6 [5.1, 28.3]                      13.3 ± 3.0 [4.3, 26.9]                        < 0.01
     Mean[Gy]                    6.6 ± 1.5 [3.0, 12.7]                       6.3 ± 1.2 [2.7, 10.9]                         < 0.01
Heart
     V20Gy[%]                    0.90 ± 1.30 [0.0, 8.0]                      0.90 ± 1.10 [0.0, 4.4]                        0.79
     Mean[Gy]                    1.4 ± 0.6 [0.5, 4.0]                        1.4 ± 0.6 [0.6, 3.3]                          0.30
Contralateral breast
     V5Gy[%]                     0.22 ± 1.29 [0.0, 10.40]                    0.46 ± 2.50 [0.0, 22.83]                      0.31
     Max[Gy]                     3.5 ± 7.3 [0.2, 43.8]                       5.0 ± 8.7 [0.2, 44.6]                         < 0.01
CI                               1.30 ± 0.09 [1.10, 1.63]                    1.24 ± 0.06 [1.14, 1.41]                      < 0.01
HI                               1.08 ± 0.02 [1.05, 1.13]                    1.07 ± 0.02 [1.05, 1.13]                      < 0.01
MU                               1205 ± 187 [775, 1650]                      1010 ± 46 [772, 1353]                         < 0.01

dose parameters for all plans stratified by the patients’         angle design with less inter-planner variability in autop-
breast sizes. In general, different breast sizes did not          lans. As it is never an easy task to convince medical
have much impact on the autoplan parameters, but for              personnel to embrace a new technology or automate
manual plans, there was slightly better ipsilateral-lung-         clinical workflow, we designed this study to compare the
sparing in smaller breast (V4Gy = 25.7 ± 3.7%) than in lar-       manual plans and in-house-developed autoplans to
ger (V4Gy = 28.1 ± 6.9%) or medium-sized (V4Gy = 29.7 ±           examine the quality and feasibility of the automated
5.8%) breast (p < 0.01).                                          plan-generating program for this large cohort study.
  For comparison purposes, the V95% of the PTV for the               The auto-planning algorithm for whole-breast irradi-
manual plan was normalized to the V95% of the PTV for             ation emulated the manual plan generation process, and
the autoplan for each patient. As shown in Fig. 4, the            required only five to 6 min to generate an optimized
DVH curves of autoplans and manual plans for the heart            plan. Compared with the available published literature
nearly overlapped. However, compared with the manual              for automated breast irradiation planning, the auto-
plans, the DVH curve of autoplans had a better                    planning algorithm developed in this study saved time in
ipsilateral-lung-sparing trend but less contralateral-            terms of plan generation and optimization. A study pub-
breast-sparing.                                                   lished in 2010 used rapid but robust two-field tangential
                                                                  IMRT planning with a speed of 9 min per plan for 53
Discussion                                                        patients with the purpose of selecting patients that might
Manual treatment planning is an iteratively trial-and-            benefit more from the deep inspiration breath hold tech-
error process that involves repeatedly adjusting beam             nique [19]. A Netherland’s study presented a single-click
angles and objective template parameters based on each            optimized autoplan software with planning time of
treatment planner’s experience, skill and knowledge.              20 min for breast and prostate and 7 min for palliative
This manual approach results in high workload and po-             vertebrae [14]. A Canadian study conducted by Purdie
tentially suboptimal plan quality due to the operator-            et al. used the Pinnacle treatment planning system to
dependent preferences and priorities of the treatment             generate optimized whole-breast irradiation treatment
planner and the physician. This study demonstrates that           plan using radio-opaque markers placed during CT
a well-designed automated treatment plan-generating al-           simulation as the only input, with the mean time of
gorithm can reduce treatment plan variability and                 7 min per plan generation [10, 11].
standardize treatment plans. This approach results from              Aside from the timesaving advantage of automated
more ideal, standardized gantry angle and collimator              treatment planning, the above-mentioned studies
Lin et al. Radiation Oncology       (2020) 15:67                                                                                            Page 6 of 9

  Fig. 3 Dose distributions in three representative patients with different breast sizes. Row (a) represents large breast size, row (b) medium breast
  size, and row (c) small breast size. Left column shows the dose distributions of autoplan, and right column the manual plan. All plans prescribed
  42.56 Gy in 16 fractions. Isodose lines were drawn with different colors

generally showed equivalent autoplan clinical acceptabil-                     autoplans in this study include the following. First, the
ity to manual plans with equivalent target dose and                           machine workload is minimized as the autoplans use
OAR constraint parameters, as well as similar treatment                       fewer MUs to achieve similar PTV coverage and OAR
delivery time, or MUs. The advantage of applying the                          sparing comparable with the manual plans overall. Other

Table 2 The constraints evaluation of ipsilateral lung, contralateral breast and heart of all plans stratified by different breast sizes. L
large breast size (n = 18), M medium breast size (n = 52), S small breast size (n = 29)
                       Manual (Mean value ± SD)                                            Auto (Mean value ± SD)
Parameter              L               M                 S                 p-value         L                 M                 S                 p-value
Ipsilateral lung
  V16Gy[%]             12.9 ± 3.8      14.5 ± 3.6        13.7 ± 3.5        0.25            13.0 ± 3.5        13.5 ± 3.3        13.1 ± 2.2        0.78
  V4Gy[%]              28.1 ± 6.9      29.7 ± 5.8        25.7 ± 3.7        < 0.01          28.8 ± 7.0        28.4 ± 5.2        26.8 ± 2.4        0.30
Contralateral breast
  V5Gy[%]              0.0 ± 0.0       0.3 ± 1.7         0.0 ± 0.4         0.46            0.0 ± 0.0         0.8 ± 3.4         0.0 ± 0.0         0.23
  Max[Gy]              2.0 ± 2.3       4.7 ± 9.6         2.2 ± 2.6         0.19            3.6 ± 4.4         6.6 ± 11.3        2.8 ± 2.9         0.13
Heart
  V20Gy[%]             0.9 ± 1.4       0.9 ± 1.3         0.9 ± 1.2         0.98            1.2 ± 1.5         0.9 ± 0.9         0.9 ± 1.1         0.42
  Mean[Gy]             1.5 ± 0.6       1.4 ± 0.6         1.3 ± 0.7         0.37            1.7 ± 0.7         1.4 ± 0.5         1.2 ± 0.5         0.03
Lin et al. Radiation Oncology     (2020) 15:67                                                                                           Page 7 of 9

 Fig. 4 Mean DVH curves with 95% confidence interval (shaded area) for the PTV, ipsilateral lung, heart and contralateral breast for autoplans
 (solid red lines) and manual plans (dashed green lines)

previously mentioned published studies used two-field                       but also find acceptable or even better solutions for a
or four-field design, with no difference in MUs between                     standardized and optimized treatment plan than a hu-
automated plans and manual plans. Second, there is a                        man could within a limited time. Four to five hypo-
slightly higher homogeneity in PTV dose coverage for                        fractionated breast cancer treatments are planned per
autoplans. The DVH analyses (Fig. 4) demonstrate a                          week at our oncology department on average. The es-
steeper slope in the average DVH curve for the PTV of                       timated working hours saved per month is therefore
the autoplans compared with the manual plans. Third,                        18 h per month. Moreover, a multi-field complex
the OAR dose parameters and target coverage are uni-                        planning task requires great experience, many calcula-
form and consistent in autoplans than in manual plans.                      tions and much trial-and-error, which may be over-
The narrow 95% confidence interval shown in the DVH                         whelming in difficult cases. Our automated planning
analyses of the 99 autoplans indicates the uniform                          algorithm achieves the best gantry angles, collimator
consistency of the dose parameters.                                         angles, MUs for each field and minimal hotspots
   The automated planning algorithm in this study is                        through iteration and optimization. This would be a
highly efficient as it uses six-field IMRT and could                        daunting task if reproduced manually. Applying an
generate an optimized plan with high conformity and                         automated planning algorithm can also reduce inter-
homogeneity in five to 6 min on a standalone work-                          planner variability between experienced and non-
station. The observed advantage of autoplans implies                        experienced treatment planners, which may contribute
that when applying a more complicated field design,                         to plan quality control and educational training.
an automated planning algorithm may not only save                             It is also interesting to note that, as shown in the DVH
much time for medical physicists and dosimetrists,                          curves in Fig. 4 and parameters in Table 1, the
Lin et al. Radiation Oncology   (2020) 15:67                                                                                     Page 8 of 9

autoplanning algorithm tends to maximize the PTV             generated by the algorithm when compared with manual
coverage without compromising it in order to spare the       plans. We conclude that a well-designed automated
lung dose, which probably results from the strict adher-     treatment planning program can improve current clin-
ence to the planning algorithm design (Fig. 1) and the       ical practice in the radiation oncology field, and argue
objective template (Additional file 1: Table S1). In real-   that automatic treatment planning will become part of
ity, human doctors and physicists might accept a trade-      the standard workflow in radiation oncology depart-
off between the PTV coverage, lung dose and other            ments at medical centers.
OAR constraints, considering a patient’s medical or clin-
ical condition, such as chronic lung disease or impaired     Supplementary information
heart function. As the criterion for plan approval is ac-    Supplementary information accompanies this paper at https://doi.org/10.
                                                             1186/s13014-020-1468-9.
cording to each clinician’s discretion, our autoplanning
algorithm has the advantage of quickly generating sev-
                                                              Additional file 1 : Table S1. The objective template for dosimetry
eral optimized plans using different objective templates,     optimization.
allowing clinicians to choose the ideal one to approve
for treatment delivery based on clinical judgement.          Abbreviations
However, physicist manual refinement of automatically        CI: Conformity index; CT: Computerized tomography; CTV: Clinical target
generated plan may be required based on individual pa-       volume; DIBH: Deep-inspiration breath-hold; DVH: Dose-volume histogram;
                                                             ESAPI: Eclipse scripting application programming interface; HI: Homogeneity
tient’s condition.                                           index; IMRT: Intensity-modulated radiation therapy; MU: Monitor unit;
   To date, automated workflow has been developed and        OAR: Organ at risk; PTV: Planning target volume; QUANTEC: Quantitative
implemented in various businesses but the pace of ap-        analysis of normal tissue effects in the clinic; RTOG: Radiation therapy
                                                             oncology group; SD: Standard deviation; TPS: Treatment planning system
plying automation into health care is not as fast as ex-
pected. In the recent two decades, the rapid progress in     Acknowledgements
computerized radiotherapy treatment planning systems         Not applicable.
has led to great interest in the possibility of automating   Authors’ contributions
radiotherapy workflow, which includes automated target       CYL, KCL and THW designed the autoplanning algorithm. CYL and TCL
delineation (auto-segmentation), automated treatment         collected the data, performed the statistical analysis and wrote the
                                                             manuscript. ACS, THW, JHJ, LCL and JAL revised and approved the final
planning, automated real-time adaptive radiotherapy and      manuscript. All authors read and approved the final manuscript.
automated quality assurance [20–25]. This study pre-
sents a well-designed fully automated planning algo-         Funding
                                                             This research is partially supported by CMUH Grant DMR-108-055.
rithm for standardized whole-breast radiotherapy
treatment plan generation with a large cohort of 99 pa-      Availability of data and materials
tients. By retrospectively analyzing and comparing the       The autoplanning code during the current study are available from CY Lin
autoplans and manual plans, we prove that this algo-         (cylin1230@gmail.com) to any reader directly on reasonable request.

rithm has great potential in clinical applicability, and     Ethics approval and consent to participate
may improve radiotherapy treatment planning workflow         Data collection and analysis were approved by the institutional review board
in efficiency and standardization. While the autoplan-       of China Medical University Hospital, Taichung, Taiwan (CMUH106-REC3–119).
ning algorithm may save much time and reduce the re-         Consent for publication
petitive planning workload of medical physicists and         Not applicable.
dosimetrists if implemented clinically, it may also help
                                                             Competing interests
medical personnel focus on tasks requiring innovation
                                                             The authors declare that they have no competing interests.
and creativity. In the future, we seek to develop a pro-
gram that automatically extracts clinical information and    Author details
                                                             1
                                                              Department of Radiation Oncology, China Medical University Hospital, China
combine it with the automated planning algorithm to
                                                             Medical University, Taichung, Taiwan. 2Department of Medicine, China
develop a personalized automated treatment planning          Medical University, Taichung, Taiwan. 3Department of Biomedical Imaging
program, thereby increasing its applicability.               and Radiological Sciences, National Yang-Ming University, Taipei, Taiwan.
                                                             4
                                                              Department of Biomedical Imaging and Radiological Science, China Medical
                                                             University, Taichung, Taiwan. 5Department of Surgery, China Medical
Conclusions                                                  University Hospital, Taichung, Taiwan.
An automated treatment-planning algorithm for stan-
                                                             Received: 18 November 2019 Accepted: 15 January 2020
dardized and optimized hypofractionated whole-breast
irradiation plan generation is developed to reduce the
radiation oncology personnel workload and improve hu-        References
man performance and productivity. Autoplans with             1. Guo F, Kuo Y, Berenson AB. Breast Cancer incidence by stage before and
                                                                 after change in screening guidelines. Am J Prev Med. 2019;56:100–8.
higher efficacy (more MU-saving) and with more uni-          2. Lin R, Tripuraneni P. Radiation therapy in early-stage invasive breast Cancer.
formity and consistency in plan parameters are                   Indian J Surg Oncol. 2011;2:101–11.
Lin et al. Radiation Oncology              (2020) 15:67                                    Page 9 of 9

3.    Yarnold J, Bentzen SM, Coles C, Haviland J. Hypofractionated whole-breast
      radiotherapy for women with early breast Cancer: myths and realities. Int J
      Radiat Oncol. 2011;79:1–9.
4.    Quantitative Analyses of Normal Tissue Effects in the Clinic (QUANTEC): An
      Introduction to the Scientific Issues - International Journal of Radiation
      Oncology • Biology • Physics. Available at: https://www.redjournal.org/
      article/S0360-3016(09)03300-8/fulltext. Accessed 11 May 2019.
5.    Rudra S, Al-Hallaq HA, Feng C, Chmura SJ, Hasan Y. Effect of RTOG breast/
      chest wall guidelines on dose-volume histogram parameters. J Appl Clin
      Med Phys Am Coll Med Phys. 2014;15:4547.
6.    Kim H, et al. Automated field-in-field (FIF) plan framework combining
      scripting application programming Interface and user-executed program for
      breast forward IMRT. Technol Cancer Res Treat. 2018;17:153303381881039.
7.    Jain P, et al. Inter-fraction motion and dosimetric consequences during breast
      intensity-modulated radiotherapy (IMRT). Radiother Oncol. 2009;90:93–8.
8.    Kestin LL, et al. Intensity modulation to improve dose uniformity with
      tangential breast radiotherapy: initial clinical experience. Int J Radiat Oncol.
      2000;48:1559–68.
9.    Ayata HB, Güden M, Ceylan C, Kücük N, Engin K. Comparison of dose
      distributions and organs at risk (OAR) doses in conventional tangential
      technique (CTT) and IMRT plans with different numbers of beam in left-
      sided breast cancer. Rep Pract Oncol Radiother. 2011;16:95–102.
10.   Purdie TG, Dinniwell RE, Letourneau D, Hill C, Sharpe MB. Automated
      planning of tangential breast intensity-modulated radiotherapy using
      heuristic optimization. Int J Radiat Oncol. 2011;81:575–83.
11.   Purdie TG, Dinniwell RE, Fyles A, Sharpe MB. Automation and intensity
      modulated radiation therapy for individualized high-quality tangent breast
      treatment plans. Int J Radiat Oncol. 2014;90:688–95.
12.   Hansen CR, et al. Automatic treatment planning improves the clinical quality of
      head and neck cancer treatment plans. Clin Transl Radiat Oncol. 2016;1:2–8.
13.   Hazell I, et al. Automatic planning of head and neck treatment plans. J Appl
      Clin Med Phys. 2016;17:272–82.
14.   De Graaf R, et al. EP-1628: single-click automatic radiotherapy treatment planning
      for breast, prostate and vertebrae. Radiother Oncol. 2016;119:S758–9.
15.   Buergy D, et al. Fully automated treatment planning of spinal metastases –
      a comparison to manual planning of volumetric modulated arc therapy for
      conventionally fractionated irradiation. Radiat Oncol. 2017;12:33.
16.   Higher Per Daily Treatment-Dose Radiation Therapy or Standard Per
      Daily Treatment Radiation Therapy in Treating Patients With Early-Stage
      Breast Cancer That Was Removed by Surgery - Tabular View -
      ClinicalTrials.gov. Available at: https://clinicaltrials.gov/ct2/show/record/
      NCT01349322. Accessed 7 May 2019.
17.   Kataria T, Sharma K, Subramani V, Karrthick KP, Bisht SS. Homogeneity index:
      an objective tool for assessment of conformal radiation treatments. J Med
      Phys Assoc Med Phys. 2012;37:207–13.
18.   Feuvret L, Noël G, Mazeron J-J, Bey P. Conformity index: a review. Int J
      Radiat Oncol Biol Phys. 2006;64:333–42.
19.   Wang W, et al. Rapid automated treatment planning process to select
      breast Cancer patients for active breathing control to achieve cardiac dose
      reduction. Int J Radiat Oncol. 2012;82:386–93.
20.   Men K, et al. Fully automatic and robust segmentation of the clinical target
      volume for radiotherapy of breast cancer using big data and deep learning.
      Phys Med. 2018;50:13–9.
21.   Liu C, et al. An automated dose tracking system for adaptive radiation
      therapy. Comput Methods Prog Biomed. 2018;154:1–8.
22.   Lamb J, et al. Online Adaptive Radiation Therapy: Implementation of a New
      Process of Care. Cureus. 9(8):e1618.
23.   Krayenbuehl J, Di Martino M, Guckenberger M, Andratschke N. Improved
      plan quality with automated radiotherapy planning for whole brain with
      hippocampus sparing: a comparison to the RTOG 0933 trial. Radiat Oncol.
      2017;12:161.
24.   Cai B, et al. Standardization and automation of quality assurance for high-
      dose-rate brachytherapy planning with application programming interface.
      Brachytherapy. 2019;18:108–114.e1.
25.   Sarkar B, et al. Standardization of volumetric modulated arc therapy-based
      frameless stereotactic technique using a multidimensional ensemble-aided
      knowledge-based planning. Med Phys. 2019;46:1953–62.

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
You can also read